Mobile CTG – Fetal Heart Rate Assessment Using Android Platform

Mobile CTG – Fetal Heart Rate Assessment Using Android Platform
Lukáˇs Zach 1 , Václav Chudáček 1 , Jakub Kuˇzílek 1 , Jiˇrí Spilka 1 , Michal Huptych 1 , Miroslav Burˇsa 1 ,
Lenka Lhotská 1
1 Department of Cybernetics, Faculty of Electrical Engineering, Czech Technical University in
Prague, Czech Republic
Abstract
Cardiotocography – measurement of fetal heart rate and
uterine contractions is the prominent source of information
about the fetal well-being in the late stages of pregnancy
and during the delivery. With the stable increase of
systematic costs of western medical systems and with the
lack of trained personnel especially in the rural areas of
BRIIC countries, telemedicine solutions are destined for
large range of users.
In the paper we describe our initial approach to proposed
solution for mobile fetal heart rate monitoring and
evaluation running on Android platform. Additionally the
application on the mobile Android device contains viewer
of the signal that enables setting of customary thresholds
levels for the analysis rules and gives user full control over
the settings of the recording device.
1. Introduction
Fetal heart activity is the prominent source of information
about fetal well being during delivery. Cardiotocography
(CTG) – recording of fetal heart rate (FHR) and uterine
contractions enables obstetricians to detect possible ongoing
fetal hypoxia which may occur even in a previously
uncomplicated pregnancy.
Cardiotocography was introduced in late 1960s and is
still the most prevalent method of intrapartum hypoxia
detection. To improve the results of cardiotocography,
the International Federation of Gynecology and Obstetrics
(FIGO) introduced general guidelines [1]. They are based
on an evaluation of macroscopic morphological FHR features
and their relation to the tocographic measurement.
Even though the guidelines have been available for more
than twenty years poor interpretation of CTG still persists
[2] with large inter-observer as well as intra-observer
assessment variations [3]. Nevertheless the FIGO guidelines
remain the only generally agreed approach to rulebased
evaluation, and therefore features we compute will
be based on the guidelines.
During the last weeks of pregnancy the frequency of pe-
riodic controls can increase up to 5 per week creating great
burden on both – mother and obstetricians. Based on our
previous research [4] dealing with automatic evaluation of
the fetal heart rate recordings during the delivery we have
developed a tool for monitoring and evaluation of the antepartum
fetal heart rate based on the FIGO guidelines.
2. Proposed mobile CTG methodology
Any FHR recording device with bluetooth module and
known data format can be prospectively connected to the
phone as a data source. After recording of at least 20
minutes the system performs signal preprocessing, including
quality assessment and then automatically computes
the features such as mean baseline of the FHR, long term
variability, and number of acceleration and decelerations.
Alarms can be raised at the patient’s side if necessary but
the data are always sent via internet (using WIFI or GPRS)
to the clinician’s database for more detailed analysis and
confirmation. A clinician can remotely ask for additional
measurement, if the first recording was inconclusive. Currently
our system works with the FHR recording module
developed within the ENIAC-MAS project funded described
in the next section. The general methodology of
the mobile CTG (mCTG) is depicted in Figure 1.
3. Data acquisition module
Currently our system works with the FHR recording
module developed within the ENIAC-MAS project funded
by the European Union. The module records fetal phonocardiogram,
which is further processed on the chip. The
data are then transferred to the mobile phone via the bluetooth
module. Prospectively any FHR recording device
with bluetooth module and known data format can be connected
into the loop via the Android enabled mobile phone
as a data source.
ISSN 0276−6574 249 Computing in Cardiology 2011;38:249−252.

Figure 1. Complete methodology envisioned for the development of the mCTG.
4. Android-based application
The Android-based application which is the focus of this
paper, runs on any Android-enabled phone with the Android
2.1 or higher – we currently use the Nexus One by
Google. The application scheme is depicted in Figure 2.
First the menu is presented to the user, where it is possible
to select between the modes of the signal acquisition.
The FHR signal can be either opened from the storage (e.g.
SD-card) – explorer activity is responsible for that. Or the
communication can go via bluetooth module – bluetooth
activity – and the interaction with the measuring device can
be initialized. In any case the data are downloaded to the
mobile phone and loaded into the application. In the following
step the data are processed in the CTG-processing
activity - which will be described in the following section.
The last step from the point of view of the Android application
design is the visualization – where either user or
the clinician can view the signals with some highlighted
features.
5. Signal processing
The processing of the signal can be divided into three
separate phases – see Figure 3. In the following subsec-
250
Figure 2. Representation of the processes within the Android
application.
tions we describe all the steps of the processing array.
5.1. Interpolation and artefact removal
In general within our application we expect FHR with
either 4Hz sampling frequency or signal with timestamps
connected to each sample – thus enabling us to resample
the possibly unevenly sampled data to the 4Hz sampling

ate.
The FHR signal almost always contains artefacts caused
by mother and fetal movements as well as artefacts caused
by transducer displacements. Based on the amount of
missing signal as well as on the amount of artefacts within
the signal, the signal quality measure is computed – which
is then used for the alarm confidence measures.
For the artefacts removal we have implemented the algorithm
proposed by Bernardes et al. [5] – all abrupt changes
in FHR were removed and replaced using cubic Hermite
spline interpolation.
5.2. Baseline assessment and ACC/DEC
detection
The computation of the baseline is based on work of
Taylor [6]. The main part of the algorithm is the use of
set of three filters, each with slightly different cut-off frequency
and order of the filter. For filtering purposes the
fast Fourier transform (FFT) was used. The signal length
N of 2 13 ought to be used since our expected length of 20minute
signal was 4800 samples. We have used slightly
different cut-off frequencies when compared to the original
work of Taylor. The maximum order of the filter was
five.
After the first filter is used and the initial baseline is established
we search for any significant signal swings on
either side of the baseline. Those with sufficient duration
and amplitudes are removed from the signal and the filtering
of the signal continues with the next filter in an iterative
way. After all the filters are used we have the final baseline
of the signal and also acceleration/deceleration positions
– positions of the significant swings above or bellow the
baseline. Number of accelerations and decelerations and
the mean of the baseline computed within this phase are
three features used in the FIGO guidelines.
Figure 3. Stages of FHR signal processing.
251
5.3. FIGO features and evaluation
The FIGO guidelines [1] use beside acceleration, deceleration
and mean of the baseline one more feature, called
long term variability (LTV) that can be computed as the
] of the distribution x(i) where
interquartile range [ 1 3
4 , 4
x(i) = � FHR 2 (i) + FHR 2 (i + 1) (1)
With four features already computed the application can
continue with assessment of the FIGO sets of rules for
antepartum monitoring – which divides signal into three
classes Normal, Suspicious, Pathological.
6. Visualization
The final part of our application is the visualization of
the results. The main screen captures are depicted in Figure
4, where the FHR signal is presented in the user interface,
that tries to mimic recent FHR visualization systems.
There is no overall information about the outcome of the
FIGO guidelines, instead each rule for each class is shown
as either fulfilled or not fulfilled.
7. Conclusions
This paper describes preliminary version of the Android
based application for the assisted living telemedicine system.
The crucial part of the system is the ability to relay
data from sensors to the hospital-based application that enables
clinicians to evaluate the signal from distance. Nevertheless
our application can also perform simple processing
and evaluation task, that could become useful e.g. for
the remote monitoring within the hospital.

Figure 4. Graphical user interface and visualization of the signal together with the indication of FIGO rules fulfillment
Acknowledgements
This work was performed as part of the project ”MAS
— Nanoelectronics for Mobile Ambient Assisted Living-
Systems” which is funded by ENIAC Joint Undertaking
and partially supported by the grant No. 7H10019 of the
Czech Ministry of Education, Youth and Sports.
References
[1] FIGO. Guidelines for the Use of Fetal Monitoring. International
Journal of Gynecology Obstetrics 1986;25:159–167.
[2] Steer PJ. Has electronic fetal heart rate monitoring made a
difference. Semin Fetal Neonatal Med Feb 2008;13(1):2–7.
[3] Blix E, Sviggum O, Koss KS, Oian P. Inter-observer variation
in assessment of 845 labour admission tests: comparison
between midwives and obstetricians in the clinical setting
and two experts. BJOG Jan 2003;110(1):1–5.
[4] Chudáček V, Spilka J, Jank˚u P, Kouck´y M, Lhotská L, Huptych
M. Automatic evaluation of intrapartum fetal heart
252
rate recordings: A comprehensive analysis of useful features.
Physiological Measurement 2011;32:1347–1360.
[5] Bernardes J, Moura C, de Sa JP, Leite LP. The Porto system
for automated cardiotocographic signal analysis. J Perinat
Med 1991;19(1-2):61–65.
[6] Taylor GM, Mires GJ, Abel EW, Tsantis S, Farrell T, Chien
PF, Liu Y. The development and validation of an algorithm
for real-time computerised fetal heart rate monitoring
in labour. BJOG Sep 2000;107(9):1130–1137.
Address for correspondence:
Václav Chudáček
Czech Technical University in Prague
Faculty of Electrical Engineering
Department of Cybernetics
Technická 2
166 27 Praha 6
Czech Republic
chudacv@fel.cvut.cz